microsoft/DeepSpeed

Pull Requests Commits

jeffra committed 3 years ago

8e4dd794

Update utils.py

jeffra committed 3 years ago

Verified e5f0f8b8

lazy load fairseq

jeffra committed 3 years ago

f4d97b1e

jeffra committed 3 years ago

c2912732

jeffra committed 3 years ago

210c6c70

add fairseq adam to zero compatible optims

jeffra committed 3 years ago

f0f2aa82

client override loss scale

jeffra committed 3 years ago

96af261f

jeffra committed 3 years ago

295ad53c

pass through attributes from self.module

jeffra committed 3 years ago

c5c8ddac

Upgrade bundler to address CVE-2020-36327

jeffra committed 3 years ago

Verified aeea85b4

contiguous gradients should be set to True by default (#1199)

samyam committed 3 years ago

Verified c9fee821

jeffra committed 3 years ago

36126122

clean up logging (#1190)

stas00 committed 3 years ago

Verified a0292398

undo noise (#1191)

stas00 committed 3 years ago

Verified bc019a53

Fix bugs about non-contiguous tensor broadcasting (#1168)

hyunwoongko committed 3 years ago

Verified 429cbc89

introduce debug utils (#1136)

stas00 committed 3 years ago

Verified c0c4ebf1

ZeRO 2+3 memory estimators (#965)

stas00 committed 3 years ago

Verified 0c1802cc

zero_to_fp32: restore persistent buffers (#1146)

stas00 committed 3 years ago

Verified df8b1f88

jeffra committed 3 years ago

0982ecca

remove torchvision dependency (#1178)

jeffra committed 3 years ago

Verified 3b689844

Avoid partitioning small activations (#1154)

tjruwase committed 3 years ago

Verified b1669c0d

add assert to make sure zero isn't enabled with 1bit-* (#1169)

jeffra committed 3 years ago

Verified 1ba3e8e3

patch in ompi local rank if local_rank isn't set (#1164)

jeffra committed 3 years ago

Verified 4c114a27

Remove hard torch dependency at install (#1166)

jeffra committed 3 years ago

Verified da1fe2f8

[Doc] Fix steps_per_print description (#1163)

tjruwase committed 3 years ago

Verified fa7921e2

Samyamr/largest partitioned params calculation fix (#1150)

samyam committed 3 years ago

Verified 4eaf9106

Fix bugs in the tutorial documentation (#1157)

hyunwoongko committed 3 years ago

Verified aa16828c

Add Windows support in README, use c++17 on Windows to support latest VC & cuda build tool (#1151)

eltonzheng committed 3 years ago

Verified 71ecf7e6

unit test for bugfix #1135 (#1148)

Shaden Smith committed 3 years ago

Verified 8e48756c

jeffra committed 3 years ago

5f16b172

Older